汉语普通话人工耳蜗使用者对声调识别的分析研究

Lexical tone perception in native-mandarin cochlear implant users

目的初步确定人工耳蜗使用者声调感知的特点,并进行分析研究。方法利用自主开发的噪声下声调识别测试材料(tone identification in noise test,TINT)在Speech Performance测试平台的控制下对20例母语为汉语普通话人工耳蜗使用者声调识别能力进行评估,获得各声调识别成绩以及声调识别混淆矩阵。采用x^2检验(chi-square test)对本组人工耳蜗使用者汉语声调识别成绩差异进行统计学分析。结果①本组人工耳蜗使用者的汉语普通话声调识别总成绩(百分制得分)分布48.75%~98.75%,平均成绩(76.38±17.29)%;②各声识别成绩平均值差异显著(x^2=2358.357,P<0.01),由高到低分别为T3(85.75%)>T4(85.25%)>T1(69.75%)>T2(64.75%),即本组人工耳蜗植入者感知T3、T4较为容易、感知T1、T2较为困难;③声调识别混淆矩阵提示T2(35.35%)、T1(30.25%)混淆度明显高于T4(14.75%)、T3(14.25%),并目各个声调之间的混淆趋势不同,T1易被混淆为T2(18.25%)、T2易被混淆为T3(24.00%)。结论时域特性明显的声调更易被人工耳蜗群体识别,因此应根据声调感知难易程度制定针对性的声调康复训练方案,综合利用听觉信息和视觉信息改善声调,特别是一声、二声的识别能力。

OBJECTIVE To explore the characteristic of lexical tone perception in native-mandarin cochlear implant users. METHODS Twenty nativemandarin cochlear implant users were recruited for this study. Tone identification in noise test（TINT） controlled by Speech Performance system were used to assess the subjects＇ tone perception ability. The analyses were performed using SPSS Statistics version 19.0. RESULTS The Tone perception scores ranged from 48.75% to 98.75%, the average score was （76.38± 17.29）%. From the best to the worst scores were T3（85.75%）,T4（85.25%）, T1（69.75%） and T2（64.75%） respectively, the difference was significant（χ2=2358.357, P〈0.01）. This result means that T3 and T4 were easier to be identified than T1 and T2. Confuse matrix results showed that T2 and T1 was easier to be confused by CI users. CONCLUSION It is easier for CI users to identify the tone with prominently temporal feature. It prompts us to design targeted rehabilitation plan.